WO1983001531A1 - Method and apparatus for positioning a transducer using embedded servo track encoding and microprocessor control - Google Patents

Abstract

A system and method for accurately positioning a transducer means (24) radially in alignment with a selected data track on a magnetic disc (22) using microprocessor control (54) and for maintaining its radial alignment. Sectored servo track (30), radially interspersed with data tracks on a magnetic disc, are uniquely encoded within groups, the code comprising the location of recorded servo signals at either a first or second position within a plurality of data frames located on the servo tracks. The recorded servo signals are detected by a transducer (24) and compared with a D.C. voltage in a comparator (44) in order to produce a series of clock pulses for a shift register (46). A predetermined sequence of pulses corresponding to possible locations for recorded servo signals within the data frame in a sector (30) is the data input to the shift register (46). The output of the shift register (46) is a binary number corresponding to the servo track or the data track with which the transducer means (24) is radially aligned. This information is processed to identify the radial location of the transducer means (24). A microprocessor (54) determined from this information the radial separation between the transducer means (24) and the selected data track and provides, if required, a signal to cause the transducer (24) to move radially.

Description

METHOD AND APPARATUS FOR POSITIONING A TRANSDUCER USIJJG EMBEDDED SERVO TRACK ENCODING AND MICROPROCESSOR CONTROL

BACKGROUND OF THE INVENTION

A. Reference to Earlier Application This application is a continuation in part of the now pending application. Serial No. .253,086, filed March 31, 1981 by the same inventors.

B. Field of the Invention The present invention relates to an improved system and technique for radially positioning a transducer relative to a rotating disc and for maintaining the transducer in the desired radial position during the rotation of the disc.

C. Prior Art In magnetic recording and reproducing systems, ro- tatable magnetic discs having a number of concentric data tracks located on them are often used. While the disc is rotating, it is essential that the magnetic transducer used to either record the data on the disc or to reproduce it from the disc remain in precise alignment with the proper data track on the disc.

A primary requirement is that the servo data take up a minimum of the disc recording surface, since that area will be unavailable for the storage of other information. Among other techniques, it has been known to intersperse servo tracks with the other information recorded on the disc. This other information is recorded in tracks radially located between the radial locations of the servo data tracks. A technique that has previously been used, and which is used with the present invention, is to provide the encoded servo data on such tracks in narrow radial sectors located on the disc surface. The remainder of the disc surface is then available for the storage of other information on data tracks. The interspersed or "embedded" servo data is encoded on the magnetic disc so that any par¬ ticular servo or data track can be uniquely identified from the signals produced by a magnetic transducer sensing the servo data. The recorded servo data must really perform two functions. It must assist the transducer positioning mechanism in accessing a selected data track and it must provide a means for enabling the transducer to follow the selected track with as little deviation as possible from the track during the rotation of the disc.

In the system described in United States Patent Nos. 4,027,338 and 4,032,984, a transducer that is radially positioned over a selected track on a magnetic disc and maintained in that radial position through the use of servo data embedded within radial sectors on tracks, as mentioned above, is described. In that system, and in other similar systems, track identification information necessary for positioning the transducer over the correct track is provided by a plurality of adjacent cells in which a signal is recorded in either one of two positions. These signal is recorded in either one of two positions. These signals actually represent reverals of magnetic flux on the disc when typical magnetic gap detectors are used. The presence of a reversal of magnetic gap detectors are used. The presence of a reversal of magnetic flux or a magnetic transition at one location within the cell is taken to represent a binary digit zero while the occurence of the reversal or transition at the second location within the cell indicates the binary digit one. The preferred code in that system is the Gray code so that 2^N tracks can be uniquely encoded when N is the number of cells provided for the track identification function in each servo track. The detection of this information provides coarse servo data that is used to position the transducer in approximate registration with a selected target track.

Additional information encoded on the servo track serves to enable the transducer to follow the proper track during the rotation of the disc. This fine positioning servo data comprises a magnetic transition SUMMARY OF THE INVENTION

In the present invention, all the servo information necessary for both the track accessing and track following functions is combined and encoded in one sequence on the servo.tracks of a magnetic disc. Importantly, the decoding scheme of the present invention allows for track identifi¬ cation to the nearest servo or data track, i.e., with a reso¬ lution of one-half of a track since the servo and data tracks are interspersed, the data tracks being located half-way between adjacent servo tracks. The track accessing function is accom¬ plished by means of a single chip microprocessor which inter¬ prets the encoded servo data and controls the movement of the transducer. Furthermore the novel encoding sequence used in the present invention enhances the stability of the transducer positioning after a preselected track has been accessed. The enhanced resolution and stability are achieved, in practical systems, using a minimum of disc space and hardware components.

In order to achieve the objects of fractional track detection, high stability and the use of a minimum amount of disc space in order to provide servo information to en¬ able a transducer positioning mechanism to access a selected track and to accurately follow that track while the disc is rotating, the present invention provides, in its preferred embodiment, sectored servo tracks interspersed with the data tracks on the magnetic disc. Each narrow radial sector has a number of frames of servo data on each servo track to uni¬ quely identify the servo track within a group of adjacent servo tracks; the number of frames being equal to at least one-half the number of tracks in the group. If the disc contains more than one group of tracks, additional means are provided to identify the particular group to which the servo track belongs. Each data frame consists of a recorded magnetic transition or reversal of magnetic flux which pro¬ vides servo information at either a first or second location within the data frame. The preferred embodiment described herein also provides an additional recorded magnetic transi¬ tion at the beginning of each data frame as a synchronization signal. The data tracks are located outside the narrow radial 2/1.

^• located at either one of two positions, the positions alternating with adjacent tracks. It will be appreciated that the system described in these patents and other similar systems that have been used use separate servo data for the track accessing and track following functions. Moreover, such systems have required the use of large numbers of logic chips to implement.

SUBSTITUTE SHEET .

sectors and have radial locations equidistant between those of the servo tracks, i.e., a transducer is correctly posi¬ tioned over a data track when it is between adjacent servo tracks. The unique code embedded in the data frames on each track includes, preferably, the requirement of a Gray code encoding, i.e., between the corresponding data frames on ad¬ jacent servo tracks, only one pair of corresponding data frames has magnetic transitions not located at the same relative locations within the two data frames. In addition, in the preferred embodiment, a stricter requirement is im¬ posed in that in the adjacent data frames from different tracks within a servo sector, the relative location of the magnetic transitions within the corresponding data frames from the different servo tracks is allowed to change only once every N successive servo tracks where N is the number of data frames per servo track. Thus, for example, if the servo tracks have four data frames, the first data frame for four adjacent tracks will be encoded one way while the next four servo tracks will be encoded the opposite way.

In order to achieve the object of track number iden¬ tification with a one-half of a track resolution, the present invention uses a dynamic comparator. In the preferred embodi¬ ment, the synchronization signals detected by the transducer are negative going pulses while the servo data signals are positive going pulses. When the automatic gain controlled or buffered signal from the transducer, therefore, has its threshold into the comparator set at a positive value, the synchronization signals do not enter the comparator, and the comparator thus compares only the positive going servo sig¬ nals from the transducer with a positive D.C. voltage equal to a fraction of the peak expected signals into the compar¬ ator. Thus, each time that the transducer passes over a radial sector of servo information on the disc, a series of pulses will be produced at the comparator output.

In the preferred embodiment four uniquely encoded data frames are used to uniquely identify eight servo tracks and their adjacent data tracks. If the transducer

-BJREA

OMPI 5.

is positioned over a servo track as it passes over the radial sector, a series of four pulses will therefore appear at the output of the comparator corresponding to the location of the signals within the four data frames of the particular servo track. If the transducer is posi¬ tioned between servo tracks, five output pulses will be produced from the dynamic comparator, two pulses being produced as the transducer passes over the data frame in which the locations of the signals differ for the corres- ponding data frames from the adjacent servo tracks. These series of four or five pulses serve as the clock for a shift register.

The data input to the shift register comprises a series of data signals that correspond temporally either to the first possible location for servo signals within the data frames or to the second possible location for servo signals within the data frames or to some predeter¬ mined combination of these. The output of the shift regi¬ ster will then be a four or five digit binary number which uniquely identifies the particular servo track within a group or a data track interspersed with the servo tracks. Thus, track identification with a resolution of one-half of a track is accomplished. If the D.C. input to the compa¬ rator is set at one-fourth of the peak level of the largest expected positive signals, this resolution is plus or minus one-fourth of a track. The use of multiple comparators having D.C. voltage inputs set at different levels can be used to produce clock signals for the shift register, and in so doing, the plus or minus one-fourth track figure can be reduced.

A unique feature of the present invention is the use of a microprocessor to control, using four main implemented algorithms, the movement of the transducer toward a selected data track at which it is desired to read or write information. The microprocessor calculates the number of tracks separating the radial position of the transducer from the radial location of the target track and updates this information each time the transducer passes over a radial sector of embedded servo infor- ation using an algorithm denominated herein as the "track accumulation interrupt" algorithm. In the preferred embodi¬ ment, if the separation between the transducer and the target track is equal to or greater than 4.00 tracks, the micropro- cessor accesses a "long seek" algorithm in which a signal is generated in order to move the transducer toward the target track at a velocity determined by the microprocessor from the track separation. The signal generated by the microprocessor is applied to a digital to analog converter so that the appro- priate analog voltage to move the transducer at the desired velocity is provided to the actuator. The value of this signal applied to the digital to analog converter is recalculated by the microprocessor using the long seek algorithm based on updated information acquired from the track accumulator inter- rupt algorithm.

If, in the preferred embodiment, the track separation between the transducer and the target track is less than 4.00 tracks, the microprocessor initially accesses a "short seek" algorithm. Using this algorithm, the microprocessor modulates a switch which, when closed, allows an analog correction voltage dependent upon the distance of the transducer from the target track to be applied to the actuator. This voltage is the same one used to maintain the transducer on a particular data track once it has been accessed and microprocessor control of the transducer position has terminated.

A "seek termination" algorithm is used by the micro¬ processor, when, during the course of a long or short seek, it determines that the transducer is close to the target track. If the transducer remains close to the target track for a measured interval of time, the seek termination algorithm permits the microprocessor control of the transducer position¬ ing to be terminated. If, however, the microprocessor determines during the course of the seek termination algorithm that in fact the transducer has overshot the target track, it causes the transducer to reverse its direction of motion. Depending on the amount of overshoot, either the long seek or short seek algorithm will be accessed by the microprocessor. The micro¬ processor continues using the four implemented algorithms, the

"BUREA OMPI . ( track accumulator interrupt algorithm, the long seek algorithm, the short seek algorithm, and the seek termination algorithm until the transducer has accessed the target track.

In the present invention, therefore, the entire track accessing function is performed using a single chip micro¬ processor and the four implemented algorithms rather than the many logic chips and hardware components that have previously been used.

The encoding of the servo tracks previously described is used to enhance the stability with which a transducer can be maintained on a particular data track. For one selected data frame the peak values of the detected servo signals oc¬ curring at times corresponding to the two possible locations for servo information are compared. This data frame is selected ^* so that the servo tracks adjacent to the selected targeted data track have different relative locations for their re¬ corded servo signals in this data frame. The comparison is performed by circuitry whose output is zero when the transducer is correctly positioned on the selected data track since the detected servo signals corresponding to the data frame are equal. If the transducer moves or drifts out of alignment by a few tracks or less producing a non-zero voltage output, the circuitry provides feedback to position the transducer over the selected data track. This circuitry is used while the microprocessor is controlling the transducer position with the short seek algorithm as well as after microprocessor control terminates. There is little chance that the trans¬ ducer will be positioned over the wrong data track as the next adjacent stable nulls are located 2N tracks on either side of the selected track as a result of the described encoding.

The novel features which are believed to be charac¬ teristic of the present invention, both as to its organi¬ zation and as to its method of operation, together with further objectives and advantages thereof, will be better understood from the following description considered in connection with the accompanying drawings, in which a pre¬ sently preferred embodiment of the invention is illustrated

O:_PI iPC* .

by way of example. It is expressly understood, however, that the description of the preferred embodiment and the drawings are for the purpose of illustration and descrip¬ tion only, and are not intended as a definition of the limits of the invention.

BRIEF DESCRIPTION OF THE DRAWINGS

FIGUI.E 1 is a simplified block diagram of a mag¬ netic disc recording and reproducing system in which the present invention is incorporated. -FIGURE 2 illustrates the two possible magnetic en¬ codings for a servo data frame on a magnetic disc using the preferred tribit encoding.

FIGURE 3 illustrates the signal from a transducer moving directly over a data frame on a servo track in which the A-phase tribit of FIGURE 2 has been encoded.

FIGURE 4 illustrates the signal from a transducer moving directly over a data frame on a servo track in which the B-phase tribit of FIGURE 2 has been encoded.

FIGURE 5 illustrates the signal from a transducer ov- ing along a data track between the two adjacent data frames on the adjacent servo tracks that are differently encoded.

FIGURE 6 is a table showing a preferred servo data encoding pattern for recording servo data on a plurality of adjacent servo disc tracks in accordance with the invention. FIGURE 7 illustrates schematically a portion of a disc having a preferred servo data encoding pattern for recording servo data on a plurality of servo disc tracks in accordance with the invention.

FIGURE 8 is a series of graphs illustrating the de- tected signals obtained for different positions of the transducer in response to the servo data encoding pattern of FIGURES 6 and 7.

FIGURE 9 is a simplified block diagram for the de¬ coding scheme used for fractional track detection and identi- fication in the present invention.

FIGURE 10 is a table showing the output of the dynamic comparator used in the decoding scheme shown in FIGURE 9.

FIGURE 11 is a table showing the output of the shift register for the different servo and data tracks within a group. 10.

FIGURE 12 shows the typical analog output for a position detector with data track zero (DT 0.0) as the targeted track.

FIGURE 13 shows the typical analog output for a position detector with data track four (DT 4.0) as the targeted track.

FIGURES 14 through 17 are flow charts showing, in simplified form, the steps of the track accumulator interrupt algorithm, the long seek algorithm, the short seek algorithm and the seek termination algorithm respectively.

11.

DESCRIPTION OF THE PREFERRED EMBODIMENT

The present invention relates to a system and method c for accessing a transducer to a selected data track on a magnetic disc and for accurately maintaining the transducer in registration with the selected track during the relative rotation of the disc with respect to the transducer. A microprocessor controlled transducer positioning servo system making use of the present invention achieves, using a minimum of disc space, high accuracy with respect to the detection and identification of the selected track and high reliability with respect to the maintaining of the transducer in alignment with the selected track, and uses a minimum of disc space and hardware to do so.

In FIGURE 1, a rotatable magnetic disc 22 which con- tains information recorded on circumferential tracks on the surface of the disc is shown schematically. This in¬ formation is recorded on the disc 22 and may be reproduced from the disc 22 by means of a transducer 24 which responds to magnetic transitions or reversals of magnetic flux occur- ring in the surface of the disc 22.

As illustrated in FIGURE 1, in order to position the transducer 24 over a selected circumferential track located on the surface of the disc 22, the transducer 24 is mounted on a carriage 26 which moves the transducer 24 radially with respect to the center of the disc 22. An actuator 28 respon¬ sive to electronic signals controls the carriage movement. It will be understood that while the preferred embodiment described herein will be described, in most instances, in terms of a system using a single disc and single transducer, the present invention is applicable to virtually any magnetic recording and reproducing system.

In the preferred embodiment of the present invention, a magnetic disc 22 has servo information encoded upon it in circumferential tracks located within radial sectors 30 of the disc surface, as shown in FIGURE 1. In the areas 32 between the radial sectors 30, working data or infor¬ mation is encoded onto the disc 22, also on circumferen¬ tial tracks. Although located in different areas of the disc it is preferred that these data tracks be located radially between the servo tracks so that a transducer is correctly radially aligned with a da .a track when it is radially equidistant from two adjacent servo track locations. The encoded servo tracks are designated, within a group of uniquely encoded tracks, by half integer numbers, such as 0.5, 1.5, 2.5, etc. with the designation ST placed before them. The data tracks, located radially between the servo tracks, are designated by integers, such as 0.0, 1.0, 2.0, etc. , with the designation DT preceeding the number. Tracks from other groups of tracks located on the disc are desig¬ nated herein by a prime, such as ST 7.5* or a double prime ST 7.5". Thus, tracks ST 7.5, ST 7.5', and 7.5" would all be identically encoded but would be distinguished from one another by means noting the particular group to which the track belonged.

As the disc 22 rotates, therefore, the transducer 24 is positioned over successive radial sectors 30 of servo information. The servo information acquired by the trans- ducer 24 as it passes over the radial sectors 30 is pro¬ cessed to produce control signals for the actuator 28, en¬ abling the carriage 26 to move the transducer 24 to a de¬ sired radial position, as will be described.

A primary objective of any transducer positioning servo system depending upon servo information encoded on the disc alongside the working data is that the servo in¬ formation take up as little space on the disc as possible, as this space is unavailable for the storage of working data. Since a rotating magnetic disc on which information is recorded on tracks may contain from less than forty to more than several hundred different data tracks, uniquely identifying each track on the disc by means of servo infor¬ mation encoded on the surface would require too much space on the disc. In the preferred embodiment of the present in- vention, therefore, servo tracks in the radial sectors on the disc are grouped into groups of eight successively po¬ sitioned adjacent tracks and uniquely identified within the group by means of a magnetic encoding. .

(

A minimum number of tracks in such a group of tracks is determined by the maximum possible radial movement of the transducer 24 between successive radial sectors 30. In order to prevent any ambiguity that could result from the tranε- ducer 24 being successively positioned on successive radial sectors 30 of tracks having the same code but from different groups, the radial distance covered by a group of uniquely encoded tracks should exceed twice the maximum possible radial movement of the transducer 24 between radial sectors 30. Also, a means for keeping track of which one of the simi¬ larly encoded tracks the transducer 24 is aligned with is re¬ quired. Preferrably, a microprocessor 54 is used to accum¬ ulate the actual track numbers as the transducer 24 moves radially, thereby providing a means for identifying the particular group in which a particular track is located.

In the preferred embodiment of the present invention in which eight successively positioned tracks are to be uniquely identified, four data frames per radial sector for each track are used. Each track is uniquely encoded by servo information in the four data frames in order to en¬ able a transducer 24 to be accessed to a selected track with¬ in the group and to follow that track accurately.

Although many methods of encoding can be used with the present invention, it is preferred to use tribit encoding, with each data frame consisting of a single tribit. The use of tribit encoding on servo tracks in connection with positioning systems for transducers was described in U.S. Patent No. 3,691,543 issued to Mueller. Each tribit consists of three possible locations for signals*. At the beginning of each tribit, a synchronization signal is recorded. The other two locations for signals are for servo information data and, in each tribit, a signal is recorded at either one or the other of these locations. Thus, each tribit can be encoded in one of two possible ways. Where the transducer 24 is a magnetic head using a gap, the signals encoded onto the servo tracks are reversals of magnetic flux occurring at specific locations. FIGURE 2 illustrates the two possible magnetic encodings for a servo ( data frame on a magnetic disc 24 in which the described tribit encoding is used. As shown in FIGURE 2, the magnetic flux φ reverses direction at the beginning of each tribit, indi¬ cated in FIGURE 2 by SI and S2, to produce a synchronization signal. The locations for possible servo information signals at which the magnetic flux would again reverse are at posi¬ tions Gl and G2 as shown in FIGURE 2. For the first possible encoding, termed the A-phase tribit, the magnetic flux re¬ verses from negative to positive at the first location Gl for possible servo information. For the other possible en¬ coding, termed the B-phase tribit, the magnetic flux reverses from negative to positive at the second location G2 for pos¬ sible servo information.

Because a flux reversal is required for each tribit at either the first or second location for servo information, the synchronization signals for adjacent tribits have the same polarity, which is opposite to the polarity of the servo information signals themselves which occur at either the first or second possible locations for servo information in each tribit.

In the preferred embodiment of the present invention, each data frame on the servo tracks consists of either a single A-phase tribit or a single B-phase tribit. FIGURES 3 and 4 illustrate, respectively, the signals resulting from a transducer 24 which is moving along a servo track directly over a data frame consisting of an A-phase tribit and a B- phase tribit. Thus, a negative synchronization pulse of amplitude E_p is produced when the transducer 24 passes over the flux reversal at position SI, while a positive pulse of amplitude E_ is produced at either position Gl or G2 when the transducer 24 passes over those positions.

If the transducer 24 is not radially positioned directly over a servo track, but instead is radially positioned equi¬ distant from two adjacent servo tracks, so that it is actually in radial alignment with a data track, the trans¬ ducer 24 will average the signals from the adjacent data frames on the two servo tracks. If the adjacent data frames are both encoded with the A-phase tribit, the transducer 24 .

( output will include a positive pulse of amplitude E. BE shown in FIGURE 3, while if the two adjacent data frames are both encoded with the B-phase tribit, the transducer 24 output signal will include a positive pulse of amplitude E_p 5 as shown in FIGURE 4. If, however, one of the adjacent data frames is encoded with the A-phase tribit and the other with the B-phase tribit, the transducer 24 output signal will have the form shown in FIGURE 5 in which positive going pulses appear at both the first and second locations for servo infor- 0 ation and have an amplitude of E_/2. The relative amplitudes of the positive going pulses in FIGURE 5 that appear at the first and second locations Gl and G2 for servo information will vary as the transducer 24 is moved closer to one servo track or the other. 5 As mentioned, each data frame within a radial sector on each of the servo tracks from a group of servo tracks is encoded with either the A-phase or B-phase tribit. The table in FIGURE 6 shows a preferred encoding of the data frames on the servo tracks within a group. As shown in the 0 table in FIGURE 6, the encoding for successive servo tracks differs within only one pair of adjacent frames. However, the coding used is more strict than the Gray code in that the encoding for adjacent data frames on adjacent tracks does not vary with the frequency that is permitted by the 5 Gray code. Instead, the variation within each data frame over the successive servo tracks is cyclical with the A- phase tribit being encoded on four adjacent tracks followed by the B-phase tribit being encoded on the following four adjacent tracks. O FIGURE 7 is a schematic representation of a portion of a radial sector 30 containing servo information. The hori¬ zontal arrows represent the direction of magnetic flux with the vertical lines in each designated servo track repre¬ senting the positions of the reversals of the magnetic flux 5 or magnetic transitions. Thus, at the beginning of each data frame, a negative magnetic transition serving as a synchroni¬ zation signal occurs on each track at the positions labeled SO, SI, S2, and S3. If the tribit encoding for a particular

OMPI Au WIPO ₍ 16.

data frame of a servo track is the B-phase tribit, the mag¬ netic transition is indicated by a vertical line in the data track occurring at the second possible location for servo information within the data frame, i.e., at the location labeled G2, such as is shown for example with respect to data frame 0 for servo track ST 0.5. Conversely, if the data frame is encoded with the A-phase tribit, the magnetic trans¬ ition occurs at the first location for servo information within the data frame, i.e., at the location labeled Gl, such as is shown for example with respect to data frame 1 of servo track ST 0.5.

In FIGURE 8, the detected signals from a magnetic transducer 24 that is radially positioned over the various servo tracks in a group and over the interspersed data tracks while the magnetic disc 22 is rotating is shown. Thus, if the transducer 24 is radially positioned directly over servo track ST 0.5 and moving along that track, as the transducer crosses the beginning of each data frame, a negative synchronization pulse of amplitude Ep occurs as shown in FIGURE 8. A positive pulse of amplitude E_p occurs in the first data frame (data frame 0) correspond¬ ing to the second possible location G2 at which a magnetic transition providing servo information could occur. For data frames 1, 2, and 3 positive pulses of amplitude E_ occur at times corresponding to the first possible loca¬ tions Gl in those data frames.

If, while the disc 22 is rotating, the transducer 24 is radially positioned in alignment with a data track and thus equidistant between two servo tracks, the output from the transducer 24 will contain positive going pulses at times cor¬ responding to both Gl and G2 for one of the four data frames of an amplitude equal to E_p/2. This signal results from the transducer 24 being positioned between adjacent servo tracks rather than directly over either one so that the contribu- tionε from each of the adjacent data tracks to the signal are averaged. At times corresponding to Gl and G2 for the other three data frames, one positive going pulse of ampli¬ tude E- per data frame occurs.

OMPI 17. ( Fractional track identification is accomplished through use of the output signals from the transducer* 24 as it passes over a radial sector 30 and produces signals corresponding to a radial location with respect to the encoded servo tracks. The output signals from the transducer 24 are processed in an automatic gain control circuit or buffer 34 shown in FIGURE 1 so that the negative going synchronization pulses have an amplitude of -4 volts. .

A block diagram of the decoding scheme used in order to identify the servo track or the data track over which the transducer 24 is positioned is shown in FIGURE 9 which shows a portion of the block diagram of FIGURE 1. The output of the automatic gain control circuit or buffer 34 shown in FIGURE 1 serves as an input to a dynamic comparator circuit 44. The other input to the dynamic comparator 44 is a D.C. voltage equal to 1/4 of the peak positive pulse that can be expected from the automatic gain control circuit 34, i.e. , one volt. This voltage determines the threshold of the dynamic comparator 44; only those pulses entering the dynamic comparator 44 that have a value greater than the comparator threshold of one volt will produce an output signal from the dynamic comparator 44. The output of the dynamic comparator 44 is digital so that pulses of fixed amplitudes are produced for all input pulses greater than one volt.

Synchronization pulses, having a negative amplitude, produce, therefore, no output from the dynamic comparator 44. Hence, the output pulses from the dynamic comparator 44 correspond to the location of the recorded servo signals within the data frames of the servo track or tracks over which the transducer is travelling. These digitized servo pulses from the dynamic comparator 44 are used as clock pulses for a shift register 46. It will be noted by refer¬ ence to FIGURES 5 or 8 that if the transducer 24 is radially positioned over the location of a data track, five clock pulses are produced at the output of the dynamic comparator 44 while if the transducer 24 is radially positioned over the location of a servo track, four clock pulses are pro- liU EA t/

OMPI 18.

( duced at the output of the dynamic comparator 44.

The table in FIGURE 10 shows the location of the clock pulses produced at the output of the dynamic comparator 44 as a function of the times HI and H2 corresponding respec- 5 tively to the locations Gl and G2 of possible servo signals within each data frame. The presence of an output from the dynamic comparator 44 at a time HI or H2 in any of the data frames is indicated in the table in FIGURE 11 by the letter C. 10 The data input or gate 48 to the shift register, in the preferred embodiment, consists of a series of pulses cor¬ responding in time to the first locations Gl for the four data frames within a radial sector 30. This series of pulses is obtained from the synchronization signals. Referring to 15 FIGURE 1, the synchronization signals from the transducer 24 are detected by the sync detector circuit 56, which produces an output only for those signals entering it which have a larger negative amplitude than -E_/2. The output of the sync detector 56 is delayed by time delay circuits 58 and 60 20 so that the output of time delay circuit 60 is a series of pulses corresponding in time to the first locations Gl for servo information within the data frames in a radial sector 30. This series of pulses is used as the Gl data input 48. The shift register circuit 46 temporarily registers 25 the presence or nonpresence of pulses at the Gl data gate 48 during the occurrence of the clock pulses. If there is a pulse at the Gl data input 48 during the occurrence of the first clock pulse from the dynamic comparator 44, a signal corresponding to a binary one is registered in the ^"30 first register QA of the shift register 46. If there is no data input occurring at the time of the first clock pulse, the information is recorded in the register QA as a binary zero. When the second clock pulse occurs, the signal stored in the first register QA is shifted to the second register 35 QB and a binary zero or one is stored in register QA, depend¬ ing upon whether there is an input corresponding to the clock pulse or not. This process continues until the last clock pulse originating from the the transducer 24 as it traverses

OMPI i ¹⁹ - the radial sector 30 has resulted in the storage of a binary zero or one in the register QA along with the other stored binary digits. At this point there is then present in the register QA through QE of the shift register a four or five digit binary number.

The output of the shift register 46 for the various servo and data tracks in a group for the Gl data input 48 into the shift register^'46 is shown in the table in FIGURE • 11. FIGURE 11 can best be understood with reference to FIGURE 10, which shows the presence or nonpreεence of the clock pulses at the times corresponding to the possible occurrences of servo signals. Thus, for example, with ref¬ erence to data track DT 0.0, there are five clock pulses so that the information registered during the occurrence of the first clock pulse will be eventually shifted to the QE register.

If, for example, the transducer 24 is positioned radially over data track DT 0.0, the first clock pulse occurs simul¬ taneously with a pulse from the Gl data gate so that a binary one is recorded and registered in QA and later shifted to QE. The second clock pulse for data track DT 0.0 occurs at a time corresponding to the second location, G2, of data frame 0 at which time there is no input into the shift register from the Gl data input 48 since that input for the preferred embodi- ent consists of pulses occurring at times corresponding to the Gl locations of the data frames only. Thus, a binary zero will be registered in register QA and eventually shifted to register QD. The last three clock pulses of data track DT 0.0 occur at times corresponding to the Gl locations of data frames 1, 2, and 3 respectively so that binary ones will be registered in registers QC, QB, and QA respectively. Hence the binary number 10111 is registered for data track DT 0.0.

For servo track ST 0.5, only four clock pulses are pro¬ duced. Hence, register QE remains empty while the data correε- ponding to the first clock pulse is shifted only over to the QD register. Since the first clock pulse for track ST 0.5 occurs at a time corresponding to the second location of a possible servo signal in data frame 0, a binary zero is .

registered in register QA and eventually shifted over to register QD. However, for the other three data frames, data frames 1, 2, and 3, clock pulses occur at times correspond¬ ing to the first locations for possible servo signals within the data frames so that binary ones are produced and regis¬ tered in registers QC, QB, and QA respectively. In the case of a servo track in which only four of the registers are used, the absence of information in the last register QE is considered as a zero. The binary number 00111 is there- fore registered for servo track ST 0.5.

Thus, as shown in FIGURE 11, a unique binary number is produced for each servo track within a group of servo tracks and for the corresponding data tracks interspersed among those servo tracks. An additional signal 52 originating from either the synchronization signals or from an additional magnetic transi¬ tion embedded upon the servo tracks of the disc 24 is used to dump the binary number in the shift register 46 into a track number decoder or memory circuit 50 where it is processed to produce a track identification number. Thus, track identi¬ fication to the nearest one half of a track is achieved. The track identification number proceeds to an interface and microprocessor unit 54, as shown in FIGURE 1, where it is com¬ pared to the number of the selected data track. In the preferred embodiment, the accuracy of the frac¬ tional track detection is plus or minus 1/4 of a track. This level of accuracy is achieved with a single dynamic compara¬ tor 44 whose threshold level is set at one-fourth the ex¬ pected peak input voltage of a positive pulse. This can be understood by considering that if the trans¬ ducer 24 is aligned radially over a data track, the input pulses to the dynamic comparator 44 (corresponding to the one data frame in which the two adjacent servo tracks have servo information signals located at different positions) have ampli- tudes of one-half of the peak expected input pulse amplitude, such as is shown in FIGURE 5. If the transducer 24 moves radially away from alignment with the data track by 1/4 of a track, one of the input pulses in the frame will increase 21.

_ in amplitude to approximately 3/4 of the peak expected input pulse amplitude and the other will decrease in amplitude to approximately 1/4 the peak expected input pulse amplitude. So long as the latter pulse is above 1/4 the value of the peak expected input pulse, an indication that the transducer 24 is-radially aligned with the data track will be given, as the output of the dynamic comparator 44 will contain the five pulses characteristic of the particular data track.

If the transducer 24 moves radially with respect to the data track slightly further towards one of the adjacent servo tracks, one pulse*ε amplitude will fall below 1/4 of the peak expected input pulse amplitude and the output of the dynamic comparator 44 will now consist of the four pulses characteristic of the servo track and the track indication given will be that of alignment with the servo track even though the trans¬ ducer 24 may at that point be as much as 1/4 track out of radial alignment with the servo track. Thus with the pre¬ ferred embodiment, the radial location of the transducer 24 may be identified to within plus or minus 1/4 of a track. Even greater accuracy is possible with the present in¬ vention through the use of additional hardware. Greater ac¬ curacy can be achieved by the use of multiple comparators mounted in parallel with different threshold voltages and by additional further signal processing equipment to combine the outputs of the comparators to produce a clock signal for the shift register.

The microprocessor unit 54 uses the track identification number to determine the number of tracks separating the trans¬ ducer 24 from the target track. In the preferred embodiment, the microprocessor 54 causes a voltage to be applied to the actuator 28 in order to move the transducer 24 at a predeter¬ mined velocity toward the target track, the velocity being dependent upon the distance spearating the transducer 24 from the target track. Thus, in the present invention, both the tachometer function, i.e., the determination of the radial distance separating the target track from the transducer 24, and the velocity scheduler function, i.e., the determination of the velocity at which the transducer 24 is moved toward the I ! target track, are accomplished with software using a single chip microprocessor rather than with hardware components utilizing several or many logic chips.

The microprocessor 54 implements four main algorithms in order to appropriately position the transducer 24 radially along the disc 22 at a selected data track. These four algo¬ rithms are denominated the "track accumulator interrupt" algo¬ rithm, the "long seek" algorithm, the "short seek" algorithm and the "seek termination" algorithm. The microprocessor contains in its memory the current radial location of the transducer 24. When it is desired to acce . a new radial location on the disc 22, the microprocessor 54 cal¬ culates the number of tracks separating the location of the trans¬ ducer 24 from the target track. Depending on this number, either long seek algorithm or the short seek algorithm is then accessed by the microprocessor 54.

If the long seek algorithm is accessed, the separation between the transducer 24 and the target track as determined by the microprocessor 54 is used in order to call from a table programmed into the memory of the microprocessor 54, a value related to the velocity with which it is desired to move the transducer 24 toward the target track. This value is appropriatel scaled and applied by the microprocessor 54 to a digital to analog converter (DAC) 82. The DAC 82 produces an analog output in response to the digital input from the microprocessor 54. This analog ouput is applied to the actuator 28 through a buffer amplifier 84, a DAC 1 switch 86 controlled by the microprocessor 54 and a power amplifier 88. The value of the signal applied to the DAC 82 is updated after the transducer 24 detects the embedded servo information from 8 successive radial sectors and calculates the radial distance travelled by the transducer 24 during the 8 sector time interval. This information is then used in the long seek algorithm in order to change the rate of movement of the transducer 24 toward the target track. When the transducer 24 is determined to be less than

1.25 tracks away from the target track, the microprocessor 54 accesses the seek termination algorithm to determine whether in fact the transducer 24 has remained close to the target track 23.

( or has overshot the target track. If it has remained close to the target track, the microprocessor 54, after 5 milli¬ seconds,* disables the track -accumulator interrupt algorithm and ceases to control the position of the transducer 24. The transducer 24 is thereafter maintained in its correct position by a comparison voltage applied through the position switch 90 from the sample and hold curcuit 70, as will be described.

In the event of overshoot, a signal is applied to the DAC 82 in order to change the direction of motion of the trans¬ ducer 24. The microprocessor 54 thereafter determines whether the transducer 24 has reversed its direction. If it has, the microprocessor 54 accesses either the long seek algorithm or the short seek algorithm, depending on the track separation, in order to move the transducer 24 toward the target track. The short seek algorithm controls the opening and closing of the position switch 90. The closing of this switch 90 applies a correction voltage produced at the output of the sample and hold circuit 70 to the actuator 28 for one milli- second intervals so long as the separation between the trans¬ ducer 24 and the target track is less than 4.00 tracks and greater than 0.50 tracks. If the separation between the trans¬ ducer 24 and the target track becomes equal to or greater than 4.00 tracks, the seek termination algorithm is accessed by the microprocessor 54 at a point where the correction for overshoot is made by reversing the direction that the transducer 24 moves.

The ability of the microprocessor 54 in the present invention to produce signals which result in the transducer 24 properly accessing the selected target track, without the use of either an external tachometer or an external velocity scheduler, can be understood with reference to the flow charts of FIGURES 14 through 17 which show, in simplified form, the basic steps of the four major pertinent algorithms implemented in the microprocessor 54: track accumulator interrupt, long seek, seek termination, and short seek.

In the flow charts of FIGURES 14 through 17, operation points within the algorithms are indicated by rectangular blocks, decision points by diamonds and transfer flags by elongated ovals. . f

During the long seek, short seek, and seek termination algorithms, updated information on the radial separation between the transducer 24 and the .target track is provided by inter¬ rupting those algorithms to calculate an updated value for the separation between the transducer 24 and the target track whenever four frames of embedded servo information are detected by the transducer 24. This is indicated by the transfer flag ^■INTERRUPT" at 100 in FIGURE 14. Referring to the flow chart for the track accumulator interrupt algorithm, FIGURE 14, after the track identification number or coarse track input (CTI) is read by the microprocessor 54 at 102, it is determined at 104 whether the track identification number or CTI read is valid or not. Since each servo or data track generates at least one signal for each data frame and may generate two signals within only one of the four data frames per sector (in the case of a data track), any identification number or CTI read indicative of a variation from such a pattern is invalid.

In the case of an invalid track number, a track iden¬ tification number is calculated by subtracting the radial distance travelled during the last sector ( STdDT) by the transducer 24 relative to the disc 22 from the last track identification number or coarse track input (LSTCTI) as indi¬ cated at 106 through 110 in FIGURE 14. The microprocessor 54 saves this new coarse track input (NEW CTI) for future use. Also, since the actual distance travelled during the sector interval (dDT) is not available in the case of an invalid track identification number, the distance travelled during the previous sector interval (LSTdDT) is used, as shown at 112. If the coarse track input (CTI) is valid, it replaces the last coarse track input (LSTCTI) and is subtracted from it in order to calculate the new distance travelled between adja¬ cent sectors (dDT) as shown at 114 through 120 in FIGURE 14.

In the case of the coarse track input (CTI) being either valid or invalid, the newly acquired distance travelled between sectors (dDT) is used to update the distance between the trans¬ ducer 24 and the target track, i.e. the distance to travel (DTT) used by the microprocessor 54 in both the long and short seek algorithms, as will be described. Depending on the distance ( 25.

between the transducer 24 and the target track, the micro¬ processor 54 determines whether double prec s.-'on or single precision.arithmetic is to be used and updates the distance to travel (DDT) accordingly as indicated at 122 through 126 of FIGURE 14. In the preferred embodiment, the faster single precision arithmetic is used when the distance between the transducer 24 and the target track is less than 16.00 tracks.

In the long seek algorithm, as will be described, signals to the DAC 82 for adjusting the velocity with which the transducer 24 moves toward the target track are applied every eight sectors, i.e., after the transducer 24 has passed •over eight radial sectors 30 of servo information and the radial distance travelled during eight sectors intervals (d8DT) has been determined. Hence, a sector counter is intially set at a value of eig and interrogated at 128. The counter is decremeted at 130 each time that the transducer 24 passes over a radial sector 30 until its count equals zero. After decrementing the sector counter, if necessary, the microprocessor 54 returns to the interrupted algorithm, as indicated by the transfer flag "RETURN FROM INTER¬ RUPT" at 132.

Initially, a command algorithm directs the microprocessor 54 to either the long seek or the short seek algorithm depending on the distance separating the transducer 24 from the target track. If the separation is less than 4.00 tracks, the micro¬ processor 54 uses the short seek algorithm, while if the distance is greater than or egual to 4.00 tracks, the microprocessor 54 uses the long seek algorithm.

A simplified flow diagram for the long seek algorithm is shown in FIGURES 15a and 15b. After an initializing step 202, the microprocessor 54 ascertains at 204 that there is currently valid sector data, i.e., that DTT has been updated and has not previously been read by the microprocessor 54 since being updated This determination, that there is valid sector data to be read, is made at various points in the algorithms and is indicated by the phrase "SECTOR" within a diamond block in the FIGURES, such as at 204 in FIGURE 15a.

OMPI _ i lO < ^*

When microprocessor 54 has determined that there is valid sector data, the microprocessor 54 at 206 sets the vari¬ able, distance to travel for the current DAC calculation (DTTCDC), equal to the distance to travel (DTT) as has been determined by the microprocessor 54 using the track accumulator interrupt algorithm. The absolute value of this quantity is also deter¬ mined. Based on the absolute value of DTTCDC, a value VSN (velocity schedule normalized) related to the velocity with which it is desired to move the transducer 24 is chosen from a table programmed into the memory of the microprocessor 54, as shown at 208 through 218 of FIGURE 15a. This value is updated only after the transducer 24 has passed eight radial sectors of servo information.

In the case where the transducer 24 and the target track are separated by more than 70.00 tracks, the maximum value of

VSN is automatically chosen without reference to the programmed table.

The determination of whether the absolute value of DTTCDC is less than 16.00 tracks at 212 is made in order to ascertain whether single precision or double precision arithmetic should be used and to provide at appropriate points in the algorithms indications or so-called flags so that the microprocessor 54 uses the desired single or double precision arithmetic.

The microprocessor 54 derives the actual value of the signal to be applied to the DAC 82, VAN, from (1) the value for VSN, (2) the radial distance travelled by the transducer 24 during the previous eight sector intervals (dβDT) and (3) the velocity of the transducer 24 at the beginning of the eight sector interval (V1N). The radial distance travelled during the previous eight sector intervals (dδDT) is calculated by sub¬ tracting the distance to travel for the current DAC calculation (DTTCDC) from the distance to travel for the last DAC calculation (DTTLDC). These operations are indicated at 220 through 232. The microprocessor then determines at 234 whether this is the first calculation being made in the long seek algorithm and proceeds accordingly. If it is the first calculation, the microprocessor 54, at 238 in FIGURE 15b, assures that the posi¬ tion switch 90, the DAC 1 switch 86 and the DAC 2 switch 92 27.

^' are all opened, and provides the selection of the target track (TGTTRK SEL) to the target track decoder circuit 80. In addition, it selects the transducer 24 to be used (TR SEL) where separate transducers are provided for each side of the disc 22. Because the transducers on either side of the disc 22 may not necessarily be aligned with each other, it is necessary when switching trans¬ ducers to synchronize at 242 the transducer's location with respec to the embedded servo information on the disc 22. When transducer are switched an additional sector interval is allowed to pass at 244 before the microprocessor 54 continues in the algorithm.

Thereafter, the microprocessor 54 closes the DAC 1 switch 86, applying the analog voltage generated in the DAC 82 to the actuator 28 through the buffer amplifier 84 and the power ampli¬ fier 88. The microprocessor 54 proceeds in the loop between 250 and 258 in FIGURE 15b for a maximum of eight sector intervals, i.e. , until the sector counter has been decremented from eight to zero, or until the absolute value of the radial separation between the transducer 24 and the target track (DTT) is less than 1.25 tracks, at which point in the latter case the micro¬ processor 54 accesses the seek termination algorithm.

In the former case, after the sector counter has been decremented to zero, it is reset back to eight at 260. If the separation between the transducer 24 and the target track is greater than 16.00 tracks, double precision arithmetic is used and the microprocessor 54 repeats the long seek algorithm beginning at 206 in FIGURE 15a in order to generate a new value of VAN to be applied to the DAC 82.

If the distance to travel is not greater than 16.00 tracks, single precision arithmetic is used and a value of VAN is generated as shown in the flow chart of FIGURE 15b between 264 and 278 in a manner similar to that used with respect to the calculation made using double precision arithmetic as pre¬ viously described and shown on the flow chart of FIGURE 15 between 216 and 230. This new value of VAN is provided to the DAC 82 at 280 and the microprocessor 54 thereafter repeats the portion of the long seek algorithm indicated on the flow chart of FIGURE 15b between 250 and 258 as has been

, 28.

Thus, it will be seen that the transducer 24 is moved toward the target track by means of the microprocessor 54 operating within the long seek algorithm and providing signals to the DAC 82 to change the velocity of the transducer 24 every eight sector intervals as it proceeds toward the target track until the distance to travel is less than 1.25 tracks, as determined at 256 on the flow chart in FIGURE 15b, at which point the microprocessor 54 accesses the seek termination algorithm. At the beginning of the seek termination algorithm, a simplified flow chart for which appears in FIGURE 16, the DAC 1 switch 86 is opened and the position switch 90 is closed. The position switch 90, as will be discussed in more detail, allows the comparison voltage appearing at the output of the sample and hold circuit 70 to be applied through the compen¬ satory amplifier circuit 94 and the power amplifier 88 to the actuator 28. When applied to the actuator 28, this analog voltage tends to move the transducer 24 toward the target track, which in turn decreases the voltage. This voltage is zero when in fact the transducer 24 is correctly positioned in alignment with the target track. The generation and characteristics of this analog voltage will be subsequently described in more detail.

Simultaneously with the opening of the DAC 1 switch 86 and the closing of the position switch 90 by the micropro- cessor 54 as shown at 302 in FIGURE 16, a 3.2 millisecond timer is started. So long as the distance to travel (DTT) is less than 1.00 but greater than -1.00 tracks (a negative distance to travel indicates that the target track has been overshot by the transducer 24), the 3.2 millisecond timer will be allowed to time out. If this occurs, the seek has essentially been completed, i.e., the target track has been accessed and the final steps of the algorithm shown in the flow chart of FIGURE 16 at 320 through 330 are performed by the microprocessor 54. These steps involve setting as new initial values for a future seek the previous destina¬ tion value for the target track (INT TRK and DST TRK) and the transducer selection (INT TR and DST TR) and also deter¬ mining whether the transducer 24 is at a track greater than 29.

256 in which case, in the preferred embodiment of the present invention, the current in the transducer 24 must be modified for a read/write operation.

Even though the microprocessor 54 has indicated that the transducer 24 is within 1.00 tracks of the target track, the transducer 24 could still be moving slightly radially. The final 5 millisecond time delay at 326 through 330 permits the transducer 24 to settle on the target track.

Since the target has now been accessed, microprocessor 54 control over the radial position of the transducer 24 ceases. The transducer 24 is maintained in correect alignment with the target track by the comparison voltage at the output of the sample and hold circuit 70 which is applied through the compensatory amplifier circuit 94, the position switch 90 and the power amplifier 88 to the actuator 28 as will be described.

If, however, the target track has been overshot by the transducer 24 by more than 1.00 tracks before the 3.2 millisecond timer has timed out, the microprocessor 54 proceeds from 310 to 312 or from 314 to 316 in FIGURE 16 and causes signals to be generated for reversing the direction of motion of the transducer 24 and moving it at the maximum possible velocity. These signals are provided to the DAC 82, the position switch 90 opened and the DAC 2 switch 92 closed as indicated at 318. The DAC 2 switch 92 functions similarly to the DAC 1 switch 86 except that a higher voltage is applied to the actuator 28 through the DAC 2 switch 92. This higher voltage is intended to overcome the mementum of the transducer 24 and reverse its direction of motion.

After it is determined that the transducer 24 has in fact reversed its direction (this is determined by noting at 334 the existence of the sector interval in which the distance travelled during a sector interval (dDT) is in the desired direction of transducer motion.)

OMPI 29/1 .

The absolute value of the distance separating the transducer 24 from the target track (DDT) is re-evaluated at 336. If this quantity is less than 4.00 tracks, the DAC 2 switch 92 is opened, the position switch 90 is closed, a 1 millisecond timer is started and the microprocessor 54 proceeds to a point within the short seek algorithm indicated by the transfer flag "WAIT 21" at 344, as will be discussed.

SUBSTITUTE SHEET 30.

(

If the absolute value of the distance to travel (DTT) is greater than or egual to 4.00 tracks, the microprocessor 54, at 340, prepares to return to the long seek algorithm, setting initial values on DTTLDC and V1N, resetting the sector counter to zero and resetting the 16.00 track indicators or flags used with respect to the selection of single or double precision arithmetic. In addition,^* the DAC 2 switch 92 is opened, all signals are cleared from the DAC 82 and the DAC 1 switch 86 is closed before the microprocessor 54 returns to the long seek algorithm as indicated by the transfer flag "WAIT 30" at 346. The short seek algorithm is accessed by the microprocessor 54 either as a result of an instruction to it during the com¬ mand algorithm or when the transducer 24 has overshot the target track by less than 4.00 tracks. Referring to FIGURE 17, a flow chart of the major steps in the short seek algorithm is shown. Blocks 402 through 414 indicate operations by the microprocessor 54 in the short seek algorithm that are similar to operations performed and already described by the micro¬ processor 54 in the long seek algorithm. Specifically, initial conditions are set and signals indicative of the particular transducer 24 and target track are outputted.

The short seek algorithm essentially modulates the posi¬ tion switch 90 open and closed so that the comparison voltage at the output of the sample and hold circuit 70 is periodically applied to the actuator 28 when the absolute value of the dis¬ tance to travel (DTT) remains less than 4.00 tracks and greater than 0.50 tracks. Initially, therefore, the position switch 90 is closed and a 1 millisecond timer is started as indicated at 416. So long as the distance to travel (DTT) remains within the limits mentioned, the 1 millisecond timer is allowed to time out and the position switch 90 remains closed. After 1 millisecond, the position switch 90 is opened as indicated at 426 and simul¬ taneously, a 4 millisecond timer is started. This timer is allowed to time out only if the absolute value of the distance to travel (DTT) remains within the range mentioned, i.e., greater than 0.50 tracks and less than or equal to 4.00 tracks. If in fact the 4 millisecond timer does time out, the position switch 90 is closed, once again applying the comparison voltage at the

OMPI IPO ( output of the sample and hold circuit 70 to the actuator 28 anά starting a 1 millisecond timer. The microprocessor 54 returns to that part of the short seek algorithm which permits the 1 millisecond timer to time out so long as the absolute value of the distance to travel (DTT) remains within the indicated range. If, during the short seek algorithm, the absolute value of the distance to travel is ascertained at 418 or at 434 to egual or exceed 4.00 tracks, it is determined whether the dis¬ tance to travel is positive or negative and the appropriate point, 312 or 316, of the seek termination algorithm concerning overshoot is accessed by the microprocessor 54 as indicated by the transfer flags "GO FWD" and "GO REV" at 442 and 444 respec¬ tively. The microprocessor 54 operates within that algorithm as has been previously described. If, on the other hand, during the seek termination algo¬ rithm, the absolute value of the distance to travel (DTT) is ascertained at 420 or at 438 to be less than or equal to 0.50 tracks, the timer is stopped (if it has been started), the posi¬ tion switch 90 is closed (if it is not closed) and the micropro- ceεsor 54 then accesses the seek termination algorithm at its initial point and proceeds through it as has been described.

Once the transducer 24 has been placed into proximity with the selected track, the transducer 24 must be maintained in alignment with that track. The same servo data, differently processed, is used for maintaining the transducer 24 in align¬ ment with the selected data track as was used to assist the transducer 24 in accessing the selected data track. This main- tenence function is enhanced by the special encoding of the data frames which has been previously discussed. In the pre- ferred embodiment, position detector circuitry compares the amplitude of the pulse at the first location for servo infor¬ mation with the amplitude at the second location in one pre¬ selected data frame and provides feedback to reposition the transducer 24 as necessary. This analog signal is provided at the output of the sample and hold circuit 70 and is applied to the actuator 28 through the compensatory amplifier circuit 94, the position switch 90 and the power amplifier 88. In the pre¬ ferred embodiment, the analog signal comparison is used not

0 P1__

SA WW1IPP0O , 32.

only to maintain the transducer 24 in position after micro¬ processor control has ceased, but also while the microprocessor 54 is operating within the short seek algorithm.

The particular data frame in which the comparison is made by the position detector circuitry depends upon the se¬ lection of the data track with which it is desired to posi¬ tion the transducer 24 in alignment. The data frame chosen for the comparison is that for which 'the servo tracks adjacent to the selected data track have recorded servo signals from the transducer 24 appearing at different locations within the data frame. Thus, as can be seen by reference to FIGURE 8, if data track DT 4.0 is the selected data track, for example, data frame 0 is chosen for the comparison.

Referring to FIGURE 1, an electrical signal repre- εenting the selected data track is sent through the micro¬ processor unit and interface 54 to a target track decoder cir¬ cuit 80. Detected synchronization signals from sync detector 56 are also sent through a data frame counter 64 to the target track decoder 80. Time delayed synchronization signals repre- εenting times HI and H2 corresponding to the Gl and G2 locations for servo signals within the data frames are applied to the tar¬ get track decoder 80 from time delays 60 and 62 respectively. In response to the applied signal representing the selected data track, target track decoder 80 produces a pair of output signals, each of which comprises a single pulse synchronized to the signal from the transducer 24. For each radial sector 30 traversed by transducer 24, each of the outputs Gl* and G2' from the target track decoder 80 comprises a single pulse precisely located with respect to the synchronization signals detected by the transducer 24 and produced at the output of the buffer 34.

The temporal locations of the Gl* and G2* pulses with respect to the synchronization signal detected by the transducer 24 and appearing at the output of the buffer 34 depends upon the target track selected. Thus, if data track DT 4.0 is the selected data track, the Gl' pulse occurs at the time corresponding to the first possible location within data frame 0 for a servo signal while the { ³³-

G2' pulse corresponds to the second possible location within data frame p for a servo signal since for data frame 0, the servo tracks adjacent to data track DT 4.0 have servo signals differently located for this data frame, as can be seen with reference to FIGURE 8.

The pulses are applied to peak detect and hold cir¬ cuits 66 and 68, the inputs of which receive the buffered signal from the transducer 24. The Gl' pulse gates the peak detect and hold circuit 66 at a time corresponding to the first possible location for a servo signal in data frame 0, so that the peak of a signal from the transducer 24 occurring at such a time will be detected and held in circuit 66. Similarly, the G2' pulse gates the peak detect and hold cir¬ cuit 68 at a time corresponding to the second possible location for a servo signal in data frame 0 so that a detected servo signal from data frame 0 occurring at a time corresponding to such location will have its peak detected and held in cir¬ cuit 68. Sample and hold circuit 70 produces an output indica¬ tive of the difference between these peaks. This output will be zero when the transducer 24 is aligned with data track DT 4.0, and will be poεitive or negative if the transducer 24 drifts away from alignment with that data track.

If the selected data track is data track DT 0.0, rather than DT 4.0, the Gl' and G2' pulses from the target track decoder 80 will be reversed, i.e., the Gl* pulse will occur at a time corresponding to the second possible location for servo information in data frame 0 while the G2* pulse will occur at a time corresponding to the first possible location for servo information in data frame 0. The output of the sample and hold circuit 70 will then be zero if the transducer 24 maintains its alignment with data track DT 0.0 and will have a non-zero value if the transducer 24 drifts out of alignment with that data track.

This output voltage is used to provide a feedback signal to the actuator 28 which moves the carriage to radially adjust the position of the transducer 24. Thus, if the trans¬ ducer 24 drifts from alignment with data track DT 4.0 in a direction towards servo track ST 4.5 the output voltage becomes

OMPI

• ^IP0 ( 34.

slightly positive. This positive voltage is fed back to the actuator 28. The compensatory amplifier circuit 94 provides high and low frequency compensation for this analog signal which is applied through the position switch 90 and the power amplifer 88 to the actuator 28 in order to cause the actuator 28 to move the carriage 26 so that the transducer 24 is moved back into alignment with data track DT 4.0. In addition, the output of sample and hold circuit 70 is digitized by fine position digitize 72 and inputted to the microprocessor 54. If a data track other than data tracks DT 0.0 or DT

4.0 as described above is selected, the output pulses from the target track decoder 80, Gl* and G2', will correspond to the first and second possible locations for servo information sig¬ nals in other data frames. For example, if data track DT 6.0 is selected, the Gl* and G2* pulses correspond to the locations of the first and second possible locations for servo information signals in data frame 2, as can be determined by reference to FIGURE 8.

As shown in FIGURE 13, when the transducer 24 is cor- rectly positioned radially over data track DT 4.0, the detected signal pulses from the transducer 24 corresponding to the first and second locations within data frame DT 0.0 are equal so that the comparison results in a zero differ¬ ence voltage and output. As the transducer 24 moves radially toward servo track ST 4.5, the difference voltage decreases since the pulse at the Gl location increases in amplitude while the pulse at the G2 location decreases in amplitude. The voltage reaches a maximum negative value when the trans¬ ducer 24 is positioned over servo track ST 4.5. Similarly, as the transducer 24 moves radially toward servo track ST 3.5, the output voltage increases reaching a maximum positive value when the transducer 24 is positioned radially in alignment with servo track ST 3.5.

It will be noted from FIGURE 13 that the output voltage is zero when the transducer 24 is aligned with data track DT 0.0 and data track DT 0.0" as well aε data track DT 4.0. However, the slope of the voltage curve is opposite when the transducer 24 is aligned with the data track DT 0.0 and DT 0.0" to what it is

_A, ( when the transducer 24 is aligned with data track DT 4.0. For data tracks DT 0.0 and DT -0.0", therefore, the feedback is regenerative so that the transducer 24 tends to moves out of alignment with these data tracks. Thus, an error signal, i.e., a positive or negative voltage at the output, is ef¬ fectively generated over eight tracks on either side of the selected data track DT 4.0; the data tracks DT 4.0* and DT 4.0" located in the groups on either side are the nearest stable nulls, i.e., positions at which the error signal is zero and the slope of the voltage is in a direction tending to align the transducer 24 to the data track.

If data track DT 0.0 is selected, the data frame 0 is still used for the comparison of amplitudes of servo infor¬ mation signals. The signals for the comparison are reversed, as has been described, in order that the slope of the output voltage curve be in the proper direction, as shown in FIGURE 12, to provide the required feedback voltage.

It will be appreciated that the distance between the stable nulls is a result of the specific encoding chosen for the servo tracks within a radial sector. If the Gray code were used, the positioning of the servo information signals encoded onto the servo tracks could vary between the first and second locations within the data frames of adjacent tracks with greater frequency, thus resulting in the stable nulls being positioned radially closer together. With the preferred method of encoding, however, the same encoding is used in the data frames of four adjacent tracks before the opposite encoding iε used for that data frame for the next four adjacent trackε. There is thus less of a chance, with the preferred embodiment, that the transducer 24 will be erroneously maintained in alignment with a non- selected data track.

The specific encoding of the servo trackε within a group in a radial sector provideε, theoretically, less capacity per data frame than a Gray code encoding would. The use of the Gray code would provide a total of 2^N unique encodings where N iε the number of data frames, while, with the presently preferred encoding, 2N unique encodings are possible. How-

OMPl

Sλ . WIPO 36. ( ever, since the maximum number of trackε in any uniquely encoded group need only exceed twice the number of trackε which the transducer can traverse between radial sectors, the limita-tion on the theoretical capacity of the present method of encoding is largely irrelevant. It should also be noted that in the system described in United States Patent Noε. 4,027,338 and 4,032,984, one data frame must be used for track following only and therefore such systems have, at best, a capacity ^:of 2^N~¹. Where N iε four aε in the preferred e bodimentε herein and in the system des¬ cribed in those patents, both systems can uniquely encode eight tracks in a group.

Since the same encoding appears on a track in each group of trackε, it iε necessary to monitor which of the similarly encoded tracks from the various groups the transducer 24 iε aligned with. Thiε can be accomplished by the microprocessor 54 accumulating the track identification numbers. The ability of the microprocessor 54 to keep an accurate count of the actual track numbers depends upon the radial rate of motion of the transducer 24 relative to the disc 22. The maximum permissible rate of motion of the transducer 24 in order to permit accumulation by the microprocessor 54 will depend upon the rate of rotation of the disc 22, the number of radial sec¬ tors 30 of servo information on the disc 22 and the number of data frames per track in each radial sector 30.

While the present invention has been described in terms of a presently preferred embodiment, other configurations and variations are within the scope of the present inven¬ tion, some of which have already been indicated. For example, the fractional track detection which is made possible by the present invention is achieved whether or not the specific method of encoding presently preferred iε used or if another encoding, such as the Gray code iε uεed, although the separa¬ tion of stable nulls would be decreased if some other encoding were used. As another example the preεent invention makeε uεe of tribit encoding becauεe thiε encoding advantageously pro¬ vides synchronization signals along with servo information signals within each data frame. However, many other methods

OMPI of encoding, such as the uni-polar dibit could be used in the present invention and the same results in terms of frac¬ tional track detection and position detector stability would still be achieved. Thus, it is intended that the claims not be limited to the specific preferred embodiment discussed herein.

Claims

( CLAIMSWe claim:

1. A servo system for accurately positioning a trans¬ ducer means radially in alignment with a selected data track on a magnetic disc having one or more groups of successively positioned circumferential servo tracks and data tracks ra¬ dially interspersed with εaid servo tracks, said track being located in a said group and said disc moving relatively to said transducer means, comprising: a plurality of adjacent data frames of recorded servo signals located on each servo track, each said data frame on a servo track located adjacent to and aligned with a corresponding data frame on adjacent servo tracks, and en¬ coded by having recorded thereon magnetic servo signals at either a first or a second location within said data frame, said corresponding first and second locations of said adja¬ cent data frames on adjacent tracks being aligned and located along the same radii, said data frames being encoded with said recorded servo signals so that said tracks within a said group are uniquely encoded, said transducer means producing detected servo signals responsive to said recorded servo signals, said detected signals being indicative of the radial position of said transducer means relative to said servo and data tracks; positioning means for moving said transducer means ra¬ dially with respect to said disc; and a first means for detecting the radial location of said transducer and for providing signals to said positioning means responsive to the detection by said transducer means of said recorded servo signals, said first means for detecting comprising: means coupled to said transducer means for producing digitized clock pulses corresponding to the times of said detected servo signals detected by said transducer; a shift register having its clock input coupled to said means for producing digitized clock pulses and having a data input which is a fixed serie ( of pulses, said series of pulses containing a pulse corresponding to each adjacent data frame on a servo track, said pulse occurring at a time corresponding to either the first or second location for recorded servo signals within said data frame, whereby said shift register registers a binary number characteristic of the radial position of said transducer relative to said servo and data tracks in a group; and means for identifying the servo or data track with which said transducer means is radially aligned from said binary number characteristic of the radial position of said transducer means and for providing a signal to said positioning means in order to move said transducer radially toward alignment with said selected data track.

-gυREAir

OMPI .

2. A servo system as in claim 1 further comprising: at least one set of recorded magnetic synchronization signals on said disc, one βychronization signal of each set being located on each servo track with said synchronization signals of each set being radially aligned with one another on said tracks, said transducer means producing detected synchronization signals responsive to said recorded syn¬ chronization signals; and wherein said first means for detecting further comprises a means for generating the fixed series of pulses used as the data input for the shift register using said detected synchroni¬ zation signals.

3. A servo system as in claims 1 or 2 further comprising: a second means for detecting the radial location of said transducer means and for providing signals to said position¬ ing means responsive to the detection by said transducer means of said recorded signals, said second means comparing the detected servo signals corresponding to said recorded servo signals at said first and second locations produced by said transducer means during a time period corresponding to one selected data frame, and producing an output signal proportional to the difference in said detected servo sig¬ nals for said one selected data frame, said data frame being selected so that said servo tracks adjacent to said selected data track have recorded servo signals in said corresponding data frames at different locations within said corresponding data frames, said output signal being provided to said posi¬ tioning means for maintaining said transducer means radially in alignment with said selected data track.

4. A servo system as in claims 2 or 3 wherein said corresponding data frames on successive servo tracks have recorded servo signals at said first or second locations within said corresponding data frames that are at the same corresponding locations over a number of successively posi¬ tioned servo tracks, said number of successively positioned servo tracks being equal to the number of adjacent data frames on a servo track.

( 5. A servo system as in claim 4 wherein said data tracks are radially located equidistant between two adjacent servo tracks and said output signal of said second means is zero if said transducer means is in radial alignment with '5 said selected data track and said second means for detecting provides negative feedback to said positioning means to radi¬ ally adjust said transducer when said transducer is out of radial alignment with said βelected data track.

€. A servo system as in claims 1, 2 or 5 wherein said means coupled to said transducer means for producing digit¬ ized clock pulses comprises at least one comparator, the first input of which is coupled to said transducer means and 5 the second input of which iε a D.C. voltage.

7. A servo system as in claim 6 wherein said data tracks are radially located equidistant between two adja¬ cent servo tracks and one comparator is used which has a D.C. voltage applied to the second input equal to 1/4 the

5 peak expected amplitude of input pulses into the first input, whereby said servo and data trackε within a group can be identified with an accuracy of plus or minus 1/4 of a track.

8. A servo system as in claim 1, 2 or 3 wherein said servo tracks are located within at least one radial sector on said disc and said data trackε are located between said sectors.

9. A servo system as in claims 2 or 3 wherein a said recorded synchronization signal is located at the begin¬ ning of each said data frame.

10. A servo system as in claim 1, 2 or 3 wherein said recorded synchronization signals and servo signals are locations of magnetic flux reversals on said disc.

-BUREA LT

OMPI . Λ. W1P0 . 4 £R_NAT 0$ ( ⁴²'

11. A method for accurately positioning a transducer means mounted on a radial positioning means into radial alignment with a selected data track on a magnetic disc moving relatively to said transducer means and having one or more groups of successively positioned circumferential servo tracks and data tracks radially interspersed with said servo tracks, said selected data track being located in one said group, said servo tracks within a group being uniquely magnetically encoded by the positioning of magneti- cally recorded servo signals at circumferential locations within a radial sector on said disc, said transducer means producing detected servo signals from said recorded servo signals on said one or two closest servo tracks as said radial sector of said disc moves relatively past said trans- ducer means, said detected servo signals from said transducer means being characteristic of the radial location of said transducer means relative to said servo and data tracks within a group, said method comprising the steps of: generating a first series of pulses corresponding to said detected servo signals in a radial sector, for application as clock pulses to a shift register; generating a fixed second series of pulses, said pulses being generated to occur simultaneously with selected possible pulses generated in the immediately pro- ceeding step, for application as input pulses to a shift register; applying said first and said fixed second series of pulses to a shift register to generate a binary number charateristic of the radial location of said transducer relative to said servo tracks in a group; decoding said binary number to identify the radial location of said transducer; comparing the identified radial location of said transducer means to the location of said selected data track and moving said transducer in response to said com¬ parison radially toward the radial location of said selected data track and repeating the above steps unless said iden¬ tified radial location of said transducer is identical to the radial location of said selected data track. (

12. A method as in claim 11 wherein each said servo track contains a plurality of adjacent data frames of recorded servo signals located on each servo track, each said data frame on a servo track located adjacent to and aligned with a corresponding data frame on adjacent servo tracks, and encoded by having recorded thereon magnetic servo signals at either a first or second location within said data frame, said corresponding first and second lo¬ cations of said adjacent data frames on adjacent tracks being aligned and located along the same radii, said data frames being encoded with said recorded servo signals so that said tracks within a said group are uniquely encoded and wherein said fixed series of pulses generated contains a pulse corresponding to each adjacent data frame on a servo track, said pulse occurring at a time corresponding to either the first or second location for recorded servo signals within said data frame.

13. A method as in claim 12 wherein said transducer is maintained in radial alignment with said selected data track by the further steps of: comparing the detected servo signals corresponding to said recorded servo signals at said first and second locations produced by said transducer means during a time period corresponding to one selected data frame, said data frame being selected so that said servo tracks adja¬ cent to said selected data track have recorded servo signals in said corresponding data frames at different locations within said corresponding data frames; moving said transducer radially toward the radial location of said selected data track if said comparison indicates that said transducer means is not radially align- ed with said selected data track. . (

14. A method as in claim 13 comprising the preliminary step of: encoding said servo tracks within a group so that said corresponding data frames on successive servo tracks have recorded servo signals at said first or second locations within said corresponding data frames that are at the same location over a number of successively positioned servo trackβ, said number of successively positioned servo tracks being equal to the number of adjacent data frames on a servo track.

15. In a system for reading or writing along generally circumferential data tracks on a magnetic disc using a radially movable transducer, the magnetic disc having servo data recorded thereon among the data tracks so that the radial position of a transducer adjacent to said disc is identifiable by the servo data detected by said transducer, the improvement in the means for causing the transducer to be placed into alignment with a selected data track comprising: of a microprocessing means for determining from said detected servo data the distance between the radial loca¬ tion of said transducer and said target track for providing a signal to cause said transducer to move toward said target track at a velocity determined by said microprocessing means from said determined distance.

16. A system as in claim 15 wherein said microprocessor comprises a single logic chip.

-gυREAfT OMPI

17. In a method for accurately positioning a transducer means into radial alignment with a selected data track on a magnetic disc having a plurality of data tracks and moving relatively to said transducer means, said disc having servo data recorded thereon among the data tracks so that the radial position of a transducer means adjacent to said disc is identi¬ fiable by the servo data detected by said transducer means, the steps of: determining from said detected servo data by microproces meanε the radial separation between the transducer means and the selected data track; and generating within said microprocessor means a signal for causing the transducer means to move at a specific velocity towa said selected data track, said velocity determined from said determined radial separation.

18. A method as in claim 17 wherein said steps of deter mining and generating are periodically repeated until said trans ducer means is moved in close proximity to said selected data track.

19. A method as in claims 17 or 18 further comprising applying said signal generated within said microprocessor means to a digital to analog converter whose output is coupled to said transducer means.

20. A method as in claims 17 or 18 wherein said signal is generated by said microprocessor means using single precision arithmetic if said determined radial separation between the trans ducer means and the selected data track is less than a designated number of tracks and by said microprocessor means using double precision arithmetic if said determined radial separation is not less than said designated number of tracks.

-BϋREA iT OMPI . ( 21. A method as in claim 11 wherein a microprocessor uses the identified radial location of said transducer means and the radial location of said selected data track in order to generate a signal to cause said transducer means to move toward said selected data track at a velocity determined by said identified radial location and the radial location of said selected data track.